Expected Assists (xA) measures the quality of chances a player creates for teammates, providing a more nuanced view of creative contribution than raw assist counts. At 1X2.TV, xA is a key feature in our AI prediction models for evaluating team attacking quality.

How xA Is Calculated

Expected Assists assigns a value to each key pass or cross based on the probability that the resulting shot will become a goal. A through ball that sets up a one-on-one with the goalkeeper receives a high xA value, while a simple square pass that leads to a long-range shot receives a low value. By summing these values, we get a player's total xA, reflecting the quality of chances they create regardless of whether teammates finish them.

xA vs Actual Assists

The difference between xA and actual assists reveals whether a player's creative output is being efficiently converted by teammates. A player with high xA but low actual assists is creating excellent chances that teammates are not finishing. Conversely, a player with low xA but many assists benefits from teammates' clinical finishing. Our models use xA as the more stable and predictive metric for evaluating creative quality.

Team-Level xA Analysis

Aggregated team xA provides insight into overall creative quality and attacking threat diversity. Teams that generate xA from multiple players and positions are more resilient to individual absences and tactical adjustments. Our models evaluate xA distribution across the squad to assess attacking depth and predict performance consistency.

Prediction Applications

High xA teams that are underperforming in terms of actual goals scored are strong regression candidates, likely to improve their scoring rate over subsequent matches. Our models identify these regression opportunities and adjust predictions accordingly, providing an edge in match outcome and goal market predictions.